[Intel-gfx] Performance regression with debug logs spammed with lots of warnings/errors

[Greg Stark]

On 30 July 2017 at 13:23, Greg Stark <stark at mit.edu> wrote:

> [drm:i915_hotplug_work_func [i915]] running encoder hotplug functions
> Jul 30 13:17:17 pixel kernel: [ 3221.929831]
> [drm:i915_hotplug_work_func [i915]] Connector HDMI-A-1 (pin 7)
> received hotplug event.
> Jul 30 13:17:17 pixel kernel: [ 3221.929850] [drm:intel_hdmi_detect
> [i915]] [CONNECTOR:57:HDMI-A-1]
> Jul 30 13:17:17 pixel kernel: [ 3221.935918] [drm:intel_get_hpd_pins
> [i915]] hotplug event received, stat 0x00800000, dig 0x00111010, pins
> 0x00000080
> Jul 30 13:17:17 pixel kernel: [ 3221.935936]
> [drm:intel_hpd_irq_handler [i915]] digital hpd port D - short

Hm. This looks like a false alarm. I disconnected and reconnected the
miniDP cable from the cheap chinese adapter to the monitor and these
immediately stopped. My guess is that on powerup the chips in the
adapter didn't quite reach a stable state and were generating these
spurious events.


-- 
greg